HauhauCS FastMTP: up to 3.02x document TG and 1.93x reasoning TG versus non-MTP — plus up to 35.2% more document TG and 21.1% more reasoning TG than standard embedded MTP.
Join the Discord for updates, roadmaps, projects, or just to chat.
Qwen3.8-27B uncensored by HauhauCS 0/465 Refusals* .
This is the Aggressive variant: direct answers, no refusal behavior, and minimal preamble on hard prompts.
Every text GGUF preserves Qwen3.8's native NextN head, and this release adds HauhauCS FastMTP: a specific acceleration sidecar qualified across the complete quant lineup at maximum native context. Vision is included through the separate BF16 projector.
Hugging Face's Hardware Compatibility widget may not recognize K_P quants. If files appear to be missing, click View variants or open Files and versions.
About
No changes to datasets or intended capabilities. This release preserves Qwen3.8-27B's text, reasoning, agentic, image, and video capabilities while applying the HauhauCS Aggressive uncensoring profile.
Pick Aggressive when you specifically want the model to get to the answer without first talking itself into compliance. For reliability-critical, specifically long-context agentic work, a Balanced release is normally the safer default when/if one is available.
BPW is the encoded tensor-payload average across the complete text model, including its embedded MTP tensors, rounded to two decimals. The projector and FastMTP sidecar work with every text quant; download the projector only for image or video input.
What are K_P quants?
K_P ("Perfect") quants are HauhauCS custom quantizations that use model-specific analysis to selectively preserve quality where it matters most. Every model gets its own optimized quantization profile.
A K_P quant effectively bumps quality up by one or two quant levels at only around 5–15% more size than the base quant. The files remain standard GGUFs and work with llama.cpp, LM Studio, and other GGUF-compatible runtimes with no special build or plugin.
Note: K_P quants may show as ? in LM Studio's quant column. This is a display issue only—the model loads and runs normally.
Specs
Dense 27B causal language model with a vision encoder
64 language-model layers
Hidden size 5,120; FFN size 17,408
248,320-token padded vocabulary
48 Gated DeltaNet layers and 16 gated-attention layers
Native embedded MTP/NextN preserved, plus the HauhauCS FastMTP 32K acceleration profile
262,144-token native context; extensible up to 1,000,000 with framework-specific configuration
HauhauCS FastMTP is the custom, variant-specific acceleration profile built for this exact Aggressive release: a compact 32K draft sidecar and per-quant serving profiles qualified for TG, acceptance, maximum native context, and VRAM.
It delivers up to 3.02x document TG and 1.93x reasoning TG versus non-MTP, plus up to 35.2% more document TG and 21.1% more reasoning TG than the standard embedded-MTP profile. The unchanged full target verifies every drafted token, so FastMTP accelerates generation without replacing the target model or changing its answers. The construction and selection methodology is exclusive to HauhauCS releases.
The benchmark ladder:
Comparison
Document TG
Reasoning TG
Scope
Standard embedded MTP vs MTP disabled
2.23x (+123.4%)
1.60x (+59.6%)
Final Q8_K_P, depth 2
HauhauCS FastMTP profile vs standard embedded MTP
+35.2%
+21.1%
Final Q8_K_P, depth 3 vs depth 2
HauhauCS FastMTP vs embedded MTP at identical depth
+11.1%
+18.2%
Final Q8_K_P, depth 3
HauhauCS FastMTP vs MTP disabled
3.02x (+202.0%)
1.93x (+93.3%)
Final Q8_K_P service
These results were measured on one RTX PRO 6000 Blackwell 96 GB per isolated lane at 204800 configured context, full CUDA offload, --no-mmap, and the official reasoning sampler. FastMTP accelerates TG; PP is reported alongside it for a complete serving comparison.
There are two acceleration paths:
Embedded MTP: use any target GGUF by itself with --spec-type draft-mtp in a current upstream llama.cpp build.
HauhauCS FastMTP: pair that same target with Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-FastMTP-32K.gguf and the HauhauCS runtime patch below.
Run HauhauCS FastMTP
The compact draft uses a standard GGUF d2t token map plus a minimal Qwen3.8 runtime consumer. Build it once. The example below uses CUDA; for ROCm/HIP or Vulkan, replace -DGGML_CUDA=ON with -DGGML_HIP=ON or -DGGML_VULKAN=ON. For CPU-only, omit the backend flag.
If draft loading reports expected 5120, 248320, got 5120, 32768, the FastMTP sidecar is correct but the executable is unpatched. Launch the freshly built ./build/bin/llama-server from this checkout.
Then serve any target quant with the one shared FastMTP sidecar:
Three-run medians for the uncached 9.8K-token document fixture and three-case means for reasoning. Every FastMTP result reproduced the corresponding embedded-MTP output hashes.
Quant
Depth
PP tok/s
Document TG
Reasoning TG
vs embedded n2, Doc / Reason
vs MTP-off, Doc / Reason
Q2_K_P
3
3351.29
213.95
145.09
+11.6% / +1.7%
2.27x / 1.48x
Q3_K_P
3
3317.16
216.15
137.99
+20.5% / +8.8%
2.54x / 1.56x
Q4_K_P
3
3204.98
187.26
123.52
+18.0% / +2.3%
2.67x / 1.71x
Q5_K_P
3
2842.05
168.29
110.50
+17.8% / +4.6%
2.61x / 1.66x
Q6_K_P
3
3081.90
156.57
103.51
+26.5% / +13.8%
2.95x / 1.91x
Q8_K_P
3
3285.86
138.18
90.07
+35.2% / +21.1%
3.02x / 1.93x
IQ2_M
3
3050.94
219.19
135.00
+13.8% / +0.4%
2.33x / 1.39x
IQ3_M
3
3269.27
204.98
128.45
+21.5% / +7.9%
2.40x / 1.45x
IQ3_XS
3
3165.75
210.64
138.34
+19.1% / +5.9%
2.38x / 1.51x
IQ4_XS
3
3445.30
211.09
135.77
+21.7% / +9.3%
2.68x / 1.66x
The full-window gate used the final scrubbed Q3_K_P and FastMTP files: 190,000 uncached prompt tokens plus 64 generated tokens completed at 1613.81 PP tok/s and 131.81 TG tok/s, with 92.0% draft acceptance and no truncation inside the configured maximum native context.
RTX 6000 Ada embedded-MTP reference speeds
Single-run reference results from the final public files at a configured max token context, full CUDA offload, --no-mmap, the official thinking sampler, and embedded MTP. The workload used an uncached 9.8K-token document-continuation prompt followed by 512 generated tokens.
Quant
PP tok/s
TG tok/s
Q2_K_P
1959.14
121.88
Q3_K_P
1944.73
112.76
Q4_K_P
1860.34
92.60
Q5_K_P
1737.51
83.25
Q6_K_P
1747.60
72.89
Q8_K_P
1827.29
59.00
IQ2_M
1884.83
121.25
IQ3_M
1867.48
108.45
IQ3_XS
1880.59
111.77
IQ4_XS
1978.46
104.25
With HauhauCS FastMTP enabled, the final Q3_K_P reached 138.37 document TG and 87.95 reasoning TG on the same Ada—23.5% and 3.9% faster than the pinned Unsloth Q3 control.
Qwen3.8 supports xhigh, medium, and low reasoning effort. Thinking and preserved reasoning are enabled by default in the official model contract.
Important:
Use --jinja for the embedded chat template.
Use the BF16 projector for Vision.
The model's native maximum is 262144.
Context length and KV precision have a large VRAM cost. Reduce context before reducing model quality if your workload does not need maximum native context.
Keep default F16 K/V on the lower tiers unless memory pressure requires otherwise.
If your llama.cpp build does not recognize the reasoning or MTP flags, update it. Older builds may still load the GGUF but will not expose the full Qwen3.8 serving path.
Turning thinking off
Qwen3.8 uses thinking mode by default. Disable it when you want shorter, faster direct responses.
K_P display: may appear as ? in UIs that do not recognize the suffix
Authenticity
Every GGUF is covered by the signed HauhauCS release manifest. Exact SHA-256 values identify byte-for-byte mirrors after renaming; canonical tensor fingerprints continue to identify HauhauCS tensors after metadata-only rewriting.
The FastMTP sidecar's exact file SHA-256 is 115e618e1f73cb50817ed5856f0551c6bf9c3d94df96f440eaca78dc63b8968b; its canonical tensor fingerprint is 49e248e799f169b6ccc6a8127b9300a95f06cf3d96a8353266f5d457e81d1c87. The public-key DER fingerprint is f7be4a2335582ab7b2e393ca1c40ce70e483f1492c0f57b8c6e05d8a7223833c.
